ABSTRACT
Is this really so? My main argument in this chapter is that although invalid intuitions may lead to errors in statistical reasoning, even statistically naïve persons hold valid statistical intuitions that can help them to solve probability problems spontaneously, given that the problems are presented in a suitable representational format. Moreover, these valid intuitions can be of much help in designing efficient training programs to teach different aspects of statistical reasoning. I begin with a broad perspective on intuitive problem solving, to show that valid statistical intuitions are just a special case of a general connection between representational format and intuitive abilities. I then illustrate this idea with two types of problems commonly used in research on statistical reasoning: Bayesian inference and the impact of sample size.1 Then I propose an associative learning explanation for why and when we have valid statistical intuitions at our disposal. Finally, I discuss the possible uses of these intuitions in training programs.
